Mr Carcharodon:I think even 20 in-lbs is marginal. I ran a quick FEA assuming 35 in-lbs, as suggested by the poster above and was getting stresses 2x the yeild value for stainless steel. I tried pushing the root radius up to .020R which helps but the limiting factor is shank below the threads. And that is where these things do seem to fail in the field. A specialty high strength stainless (say nitronic 50?) might be a better choice. Kind of a marginal design.
I just did a basic classic analysis and I came up with relatively high stresses, but not close to twice the yield strength of SS.
At 35 in*lbf I calculated a combined principal stress (torsional shear + tensile) of 18,876psi. Compared to yield strength for 304SS (35,000 psi), it is kind of high IMO for this type of part.
You may notice that in the drawing I requested a quote with 17-4PH stainless steel. The Navy and other marine industries love that SS for its strength and corrosion resistance.
I would think if the stresses due to torque were twice yield we would be seen far more documented failures. In the 70s I dough many technician used a torque wrench, I know I didnt.
IMHO, I am thinking that there may be some conservative assumption in your model.
I am just curious, what FEA software are you using?
If you PM me with you email address, I will be glad to send you a PDF copy of my calculations. Actually, I can send you the Mathcad file if you like.
In any case at this moment the only price quote I got for SS reproduction of the part is too high. I am pursuing locating parts from one of the cloning manufacturers.
In the mean time I think I may switch my wifes D400 to an Aqua Lung Titan first stage (same as the venerable Conshelf with better LP port location for better hose routing).
I normally use a Royal Aqua Master (Phoenix conversion) double hose regulator (same first stage mechanism as the Conshelf). There are no possibilities of a connection failure in this regulator, since there are no connection or LP hoses to fail well that was before I added an octopus, etc. I always think it is kind of funny how the extra safety equipment will actually decrease the "meantime between failure" (MTBF).